AI Industry in Flux: Advertising, Browsers, and Safety Concerns
A synthesis of recent reports reveals multiple concurrent developments in the artificial intelligence industry, including research on chatbot advertising influence, the launch of a new AI-integrated browser, low consumer subscription uptake, and ongoing tensions between commercial priorities and safety considerations.
Chatbots could be more effective than social media algorithms at persuasion due to their ability to solicit information directly and adapt to user emotions.
Research on Chatbot Advertising Influence
A study published in an Association for Computing Machinery journal examined the potential for AI chatbots to influence user decisions through covert advertising. Researchers developed a chatbot that inserted undisclosed product recommendations into conversations. The study involved 179 participants who completed tasks using one of three chatbot versions: a standard chatbot, one with undisclosed ads, and one with labeled sponsored suggestions.
Key findings from the study include:
- The ad-infused chatbot suggested products based on conversation context, such as recommending a calorie-tracking app when a user requested a diet plan.
- Half of the participants who received sponsored but disclosed ads did not recognize the advertising language.
- Participants rated ad-infused responses as more friendly and helpful, despite the chatbot performing 3% to 4% worse on task completion.
The study's authors noted that chatbots can infer personal data from routine queries, enabling detailed user profiling. They also suggested that chatbots could be more effective than social media algorithms at persuasion due to their ability to solicit information directly and adapt to user emotions.
OpenAI's Atlas Browser Launch
OpenAI has introduced Atlas, a web browser that integrates its ChatGPT technology. The browser is currently available for Apple computers. OpenAI CEO Sam Altman stated that artificial intelligence presents an opportunity to redefine the browser experience.
Browser Features
Atlas includes an "agentic mode" feature enabling the browser to perform actions on behalf of the user. Demonstrated capabilities include:
- Analyzing an online recipe
- Calculating required ingredient quantities for a specified number of diners
- Facilitating online purchase of those ingredients
The browser can interact with user services such as email and Google Docs and can retain "browser memories" from visited websites.
Data Privacy Considerations
The integration of ChatGPT within Atlas has prompted discussion regarding user data privacy.
Anil Dash, a technology entrepreneur, suggested that the system might transmit more information to OpenAI than it provides to the user.
Lena Cohen, a technologist at the Electronic Frontier Foundation, raised concerns about AI browsers operating in agentic mode, noting that users might transfer more control to OpenAI than initially perceived and that managing data stored on OpenAI's servers could become complex.
OpenAI has stated that its default setting does not utilize information accessed via Atlas for training its AI models; however, users have the option to consent to this use.
Security Concerns
Cohen identified "prompt injections" as a potential security risk in AI browsers. Prompt injections are malicious instructions embedded within web pages that an AI agent could be induced to execute. Examples provided include an agent being directed to purchase an unintended product or to disclose credit card information.
OpenAI acknowledges prompt injection as an unresolved issue and reports ongoing efforts to train its models to disregard such instructions.
Chirag Shah, a professor at the University of Washington's Information School, commented on the rapid development of AI with minimal regulatory frameworks.
AI Subscription Market
As of February 2025, approximately 3% of US households had an AI subscription, according to the Bank of America Institute. This figure represents a 10% increase from the previous year.
Key market details:
- OpenAI's ChatGPT Plus and Google's Gemini each cost approximately $20 per month.
- Free versions of AI chatbots often limit the number of messages or provide access to weaker models.
- Paid subscribers can send more messages, use more advanced models, and access features like custom versions of ChatGPT.
Nick Turley, head of ChatGPT at OpenAI, stated that the company aims to keep a free version available but does not guarantee a free version of future, more advanced models.
Sekoul Krastev of the Decision Lab compared AI subscription growth to the early days of streaming services, predicting a sharp increase once the norm is established.
Industry Shift Towards Advertising
Reports indicate a shift in the AI industry towards advertising-based monetization. OpenAI launched its ChatGPT Search feature in late 2024 and its Atlas browser in October 2025. Sam Altman, who previously described combining ads and AI as "unsettling," now states that ads can be deployed in AI applications while maintaining trust.
Other company implementations include:
- AI search company Perplexity began experimenting with ads in 2024.
- Microsoft introduced ads to its Copilot AI.
- Google's AI Mode for search increasingly includes ads.
- Amazon's Rufus chatbot includes ads.
Security experts and data scientists have noted that AI has the potential to influence users' thoughts, spending habits, and beliefs through active dialogue, which differs from traditional web search advertising methods.
Research cited in source materials indicates:
- A December 2023 meta-analysis of 121 randomized trials found AI models to be comparable to humans in shifting perceptions, attitudes, and behaviors.
- A more recent meta-analysis of eight studies concluded no significant overall difference in persuasive performance between large language models and humans.
Safety Concerns and Researcher Departures
Several AI safety researchers have recently resigned, citing concerns that companies are prioritizing profits over safety.
OpenAI researcher Zoë Hitzig warned that integrating advertisements into AI chatbot interactions could lead to manipulation. OpenAI has stated that ads do not influence ChatGPT's responses.
Anthropic safety researcher Mrinank Sharma resigned, expressing concerns about aligning actions with values within the industry.
Personnel Changes
Fidji Simo, known for developing Facebook's advertising business, joined OpenAI in 2024. OpenAI also terminated executive Ryan Beiermeister for sexual discrimination; reports suggest Beiermeister had opposed the introduction of adult content.
International AI Safety Report
The International AI Safety Report 2026, which details risks including faulty automation and misinformation and proposes regulatory frameworks, received endorsement from 60 countries.
The US and UK governments did not sign the report.